Continue in 2 seconds

The Best Actionable Decisions

  • May 02 2003, 1:00am EDT

Business intelligence as we know it today grew out of the need to pull information together from multiple disparate systems as well as retain a historic perspective of how business operations change over time. Many organizations believe that decisions and operational adjustments made today should be influenced by observing trends and patterns from the past. Making the best actionable decisions requires the ability to draw upon information and solutions that make up one’s business intelligence framework.

Business Intelligence Heritage

The first form of business intelligence solutions surfaced as sales information systems, dating back to the seventies. As these systems matured and information from other functions within the organization funneled into them, the term “data warehouse” began surfacing. Consequently, organizations began building applications and solutions that took advantage of this consolidated view of information. As organizations began to evolve these solutions, the noun – data warehouse – shifted to the verb – data warehousing – to include both the physical structure and the applications sitting on top. Bundling both technology and business solutions together using such a common expression presented confusion to many business professionals.

The business application sitting on top of the enabling technology was clearly the most dominate of the two and, rightly so. Primarily because how individuals interact and use data warehouse solutions is the most important characteristic. How the solution is being used to solve real problems is essential. Better decisions and higher intelligence go hand in hand. Business intelligence surfaced as an updated way to describe these solutions and enabling framework. This term quickly took hold.

Today, the evolving definition for business intelligence can best be described as a framework that encompasses both solutions and enabling technology components designed to enhance the decision-making process within an organization. How one uses the framework is very important. Does the framework allow business users to make the best decisions? The discipline of obtaining “the best” answers through optimization is one of the core building blocks associated with business intelligence.

Optimization Basics

Does the insight obtained from your business intelligence framework answer questions? Are the answers that you are getting the best? Today, many organizations are forced to make do with less. Not making the right decisions or addressing questions that surface around a given business problem with suboptimal answers can mean the difference between success and failure.

Today’s business problems can be broken down into three main components – goals, decision variables and constraints. Goals and constraints are not always controllable. For example, goals may be extremely aggressive due to competitive pressure, while constraints limit the amount of resource (cash, people, assets, etc.) one can allocate to a give business scenario or problem.

Fortunately, one can control and establish values for decision variables. Selecting the right value for decision variables is the primary objective of optimization. Figure 1 breaks down the components of an optimization problem. Notice the outcome is not just “the best” solution, but rather the best actionable solution. There is a big difference between the two. Reality plays a big part in selecting the appropriate outcome. This has to be aligned with realistic capabilities of the organization. Additionally, the cost and other trade-offs associated with the best solution need to be taken into account. For example, if limited funds are available for investing and one wants to allocate funds across different financial instruments, the plan with the highest return will look the most attractive. However, when taking risk into account, the attractiveness changes. Given a specific tolerance for risk and given limited funds, how should one allocate money across multiple financial instruments with different returns and risk profiles? This is a classic business scenario.

Figure 1: Components of an Optimization Problem

Solution Approach

One can attempt to solve optimization problems using three different techniques. Selecting the right technique depends on the characteristics of the problem. Simple techniques may work well for problems with one or two variables and a small number of constraints, but given the nature of today’s business scenarios with multiple variables and constraints, simple techniques fall short of producing the right answers. The basic techniques include:

  • Iterative/Ad Hoc – This approach is risky and leaves a lot to chance. Using gut instinct to select values in some cases is counterintuitive. One is left with the nagging suspicion that things can be better.
  • Exhaustive – This approach works for business problems that have few variables and constraints. This type of problem can be modeled in such a way that every single solution can be analyzed for effectiveness. For example, if one has four decision variables and each decision variable has 10 possibilities one could calculate 104 (10,000) possibilities. If it took each possibility three seconds to run through the modeled calculation, it would take a little more then eight hours to run through every single possibility. This is an acceptable investment in time for scenarios dealing with manageable number of variables. However, things can get out of hand quickly. If one doubled the variables, the same problem would take over 9.5 years to run through every possible solution.
  • Solution Discovery - This approach leverages various methods and techniques that have been evolving for some time. These methods are capable of navigating the solution space and quickly determining what looks good and what does not by focusing on the extreme points. Packages that support this type of sophisticated decision making are readily available. The difficulty and challenge with these solutions is knowing how to integrate them into the business intelligence framework in such a way that the end users do not get bogged down with the statistical or mathematical detail behind them.

Solution discovery can be used for solving multiple problem types. Examples of these include price/revenue/profit optimization, portfolio optimization, scheduling/logistical optimization and product mix scenarios. These all deal with multiple decision variables and constraints. Determining the best actionable decision or approach for solving these business scenarios can mean the difference between success and failure. The business intelligence framework can be used to design and support this type of solution, especially if structured in a way that insulates the business user from the mechanics.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access